Community Partitioning Combining Topological Structure and Multi-attribute Characteristics

Abstract

In complex networks, the community division of nodes is often based on the topology of the network. In contrast, in real networks, the attributes of nodes themselves also affect the relationships between and within communities. Due to the multi-attribute diversity of social network platforms, it is not accurate to divide network users only from network topology. Therefore, a community division algorithm is proposed based on tag propagation algorithm combining node topology and attribute characteristics. And randomness of label propagation algorithm and instability, so the degree of similarity between nodes will combine influence and to optimize the spread of the label the initial stage, reduce the randomness, and combining the network users interested in tag attributes and the user activity to improve the communication process, makes the division of the community structure of the community more and more obvious attribute. To prove the effectiveness of the proposed method, we compare single attribute, multi-attribute, and the strength of community structure on the real microblog user datasets.